RESEARCH ARTICLE Open Access

Vitamin D attenuates inflammation, fattyinfiltration, and cartilage loss in the kneeof hyperlipidemic microswineVikrant Rai1, Nicholas E. Dietz2, Matthew F. Dilisio3,4, Mohamed M. Radwan1 and Devendra K. Agrawal1*

Abstract

Background: Osteoarthritis (OA) of the knee joint is a degenerative process resulting in cartilage loss. Recentevidence suggests that OA is not merely a disease of cartilage but a disease of the entire knee joint and thatinflammation may play an important role. OA has been associated with vitamin D deficiency. Vitamin D as animmunomodulator and anti-inflammatory agent may attenuate inflammation in the knee. The aim of this studywas to assess the anti-inflammatory effect of vitamin D on inflammation in the knee.

Methods: This study was conducted with 13 microswine on a high cholesterol diet categorized into three groupsof vitamin D-deficient, vitamin D-sufficient, and vitamin D supplementation. After 1 year, microswine were killed,and their knee joint tissues were harvested. Histological and immunofluorescence studies were carried out on thetissue specimens to evaluate the effect of vitamin D status.

Results: Histological and immunofluorescence studies of the knee joint tissues showed (1) increased inflammationin the knee joint tissues, (2) fatty infiltration in quadriceps muscle, patellar tendon, and collateral ligaments, and(3) chondrocyte clustering in the vitamin D-deficient and vitamin D-sufficient groups compared with the vitamin Dsupplementation group. Architectural distortion of the quadriceps muscle, patellar tendon, and collateral ligamentswas also seen in the areas of inflammatory foci and fatty infiltration in the vitamin D-deficient group.

Conclusions: Decreased inflammation and fatty infiltration in the vitamin D supplementation group suggest thepotential role of vitamin D in attenuating inflammation and fatty infiltration as well as in protecting the architectureof the tissue in the knee joint.

Keywords: Osteoarthritis, Vitamin D deficiency, Cartilage loss, Inflammation, Fatty infiltration, Vitamin Dsupplementation, Inflammation attenuation

BackgroundInflammation of the joint accompanied with pain andswelling is a major problem leading to joint damage. Thisinflammation is accompanied with loss of joint flexibilityand range of motion, resulting in physical disability andlimitation of movement. Among the hand, hip, and kneejoints, knee joint arthritis is the most common form ofarthritis [1]. Osteoarthritis (OA) of the knee joint is a de-generative joint disease that is the foremost cause of dis-ability in the elderly population in the United States [2].

OA is characterized by an imbalance between degenera-tive and regenerative processes, resulting in cartilage loss[3]. Although the pathogenesis of cartilage degeneration isstill unclear, there is evidence which suggests that cartilagedamage is due to periarticular bone resorption andsclerosis during subchondral bone remodeling. Thissubchondral bone remodeling response may be slowed bylow levels of vitamin D, resulting in bone thickening,osteophyte formation, and resultant cartilage damage[4, 5]. Therefore, low levels of vitamin D may affect struc-tural changes of the knee joint [3]. OA affects all the kneejoint tissues, including cartilage, muscle [2, 6], tendon, liga-ment, and subchondral bone, and these changes play animportant role in the pathogenesis of OA [7]. Vitamin D

* Correspondence: dkagr@creighton.edu1Department of Clinical and Translational Science, Creighton UniversitySchool of Medicine, CRISS II Room 510, 2500 California Plaza, Omaha, NE68178, USAFull list of author information is available at the end of the article

© 2016 The Author(s). Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Rai et al. Arthritis Research & Therapy (2016) 18:203 DOI 10.1186/s13075-016-1099-6

influences the state of these articular structures by actingthrough vitamin D receptors (VDRs). VDR polymorphismhas been associated with OA, and therefore vitamin D mayplay an important role in OA pathogenesis [8–10].Vitamin D deficiency is common worldwide [11].

Vitamin D deficiency has been associated with manymusculoskeletal diseases, such as muscle weakness, rick-ets, osteomalacia, osteopenia, and osteoporosis, as wellas increased risk of fracture and muscle weakness [12].The important role of vitamin D in bone mineralization,remodeling, and maintenance is well known, but the roleof vitamin D in the pathogenesis of OA is yet to bedefined [4]. Low levels of vitamin D are associated withprogression and increased prevalence of OA [13–16].Many studies support the beneficial role of vitamin D inOA [17, 18], but this is controversial [19, 20]. Low levelsof vitamin D have also been associated with an increasedincidence of inflammation [21, 22]. Recent evidencesuggests a potential role of inflammation in OA pathogen-esis [23, 24], and vitamin D as an immunomodulatory andanti-inflammatory agent may attenuate inflammation inthe knee.Macrophages are potent modulators of inflammation

and, as sentinels of the innate immune system, are involvedin the inflammatory response. OA is a wear-and-teardisease, and wear particles also stimulate a macrophage re-sponse [25]. Macrophages and macrophage-produced cyto-kines play a potential role in the pathogenesis of OA [26].Thus, inflammatory mediators or markers expressed onmacrophages may play a role in the pathogenesis of OA.Triggering receptor expressed on myeloid cells (TREM)-1is a recently discovered amplifier of inflammation ex-pressed on monocytes and/or macrophages and neutro-phils, and TREM-2, an anti-inflammatory marker secretedfrom macrophages and dendritic and microglial cells,plays a key role in many inflammatory diseases [27–29].TREM-1 plays a potential role in the pathogenesis ofrheumatoid arthritis [30]. However, the role of TREM-1and TREM-2 in OA is largely unknown. Further, earlyinnate response due to trauma to the joint results insecretion of adiponectin and leptin by adipose tissue[7, 31–33]. The effect of vitamin D status on release ofthese adipokines in inflamed knee joints is largely un-known. Because vitamin D is an immunomodulatory andanti-inflammatory agent, vitamin D supplementation mayaffect the expression of TREM-1, TREM-2, adiponectin,and leptin, but this association is currently not well defined.Vitamin D deficiency and decreased expression of

VDR are associated with increased inflammation ofepicardial fat, and vitamin D supplementation reducesthis inflammation [34]. Further, hyperlipidemia and highfructose are also instigators of inflammation [34, 35].While studying the effect of vitamin D status on thedevelopment of atherosclerotic lesions in the coronary

arteries of swine fed a high-cholesterol and high-fat diet,we observed increased inflammation in the knee of thevitamin D-deficient swine. Therefore, we planned to evalu-ate the effect of vitamin D status (deficient, sufficient, andsupplemented) on inflammation, TREMs, adiponectin, lep-tin, and change in the histology of the knee joint tissues inthese microswine. We hypothesized that vitamin D supple-mentation should decrease inflammation in the knee jointtissue. The purpose of this study was to evaluate the bene-ficial immunomodulatory and anti-inflammatory roles ofvitamin D in attenuating inflammation in the knee.

MethodsPorcine modelThis study was conducted using Yucatan female micro-swine [36, 37]. In this study, 11 microswine were fed apelleted, high-cholesterol diet (Teklad Miniswine Diet;Envigo, Indianapolis, IN, USA) with 17.4 % protein,45.2 % carbohydrate, and 10.0 % fat by weight. Of thetotal kilocalories derived from the diet, 20.4 % of thekilocalories were from protein, 53.1 % were from carbo-hydrate, and 26.4 % were from fat. Two swine were fed apelleted, high-fructose diet (1.5 IU/g vitamin D, 17.6 %fructose, 4 % cholesterol, 1.5 % sodium cholate; TekladMiniswine Diet TD-150253) with 17.4 % protein, 46.8 %carbohydrate, and 10.0 % fat by weight. Of the totalkilocalories derived from the diet, 20.1 % of the kilo-calories were from protein, 54.0 % were from carbo-hydrate, and 26.0 % were from fat. The pelleted dietwas either vitamin D-deficient (TD-150251), vitamin D-sufficient (TD-150250; 1.5 IU/g vitamin D, 4.5 % hydroge-nated vegetable oil, 4 % cholesterol, 1.5 % sodium cholate)with 1500 IU/day of vitamin D3, or supplemented (TD-150252; 5 IU/g vitamin D, 4.5 % hydrogenated vegetableoil, 4 % cholesterol, 1.5 % sodium cholate) with 5000 IU/day of vitamin D3. On the basis of our previous experienceand other studies, vitamin D3 supplementation with1500 IU/day for the vitamin D-sufficient group or5000 IU/day for the vitamin D-supplemented group wasadded to achieve normal (21–29 ng/ml) or supplementedserum levels of vitamin D in microswine [33, 38, 39].Serum 25(OH)D levels were measured regularly. Bloodserum levels of vitamin D were measured and recorded.The swine were placed into three respective categoriesbased upon their 25(OH)D levels. The 25(OH)D-level pa-rameters used for the classification were as follows: Levelsin vitamin D-deficient (VDDef) swine were ≤20 ng/ml;levels in vitamin D-sufficient (VDSuff ) swine were30–44 ng/ml; and levels in vitamin D-supplemented(VDSupp) swine were >44 ng/ml. Five swine were in theVDDef group, 5 were in the VDSuff group, and 3 were inthe VDSupp group, for a total of 13 swine in this study.After 1 year of being fed the diet, the microswine werekilled, and knee joint tissues were collected.

Rai et al. Arthritis Research & Therapy (2016) 18:203 Page 2 of 17

Tissue acquisitionThe knee joint tissues (suprapatellar fat, quadricepsmuscle, articular cartilage of tibial tuberosity, patellartendon, collateral ligaments, medial and lateral menisci,infrapatellar fat pad, and synovial membrane) were har-vested quickly after the swine were killed. Then the tis-sues were transported to a laboratory and fixed in 4 %formalin buffer for 24 h. Each specimen was transverselysectioned at 2 mm, processed in a Tissue-Tek VIP tissueprocessor (Sakura Finetek, Torrance, CA, USA), andembedded in paraffin. Thin sections (5 μm) were cut usinga microtome (Leica Biosystems, Buffalo Grove, IL, USA)and subsequently placed on slides for hematoxylin andeosin (H&E) staining and immunofluorescence evaluation.

Hematoxylin and eosin staining of specimensAll the knee joint tissues were stained with H&E follow-ing the manufacturer’s standard protocol (NewcomerSupply, Middleton, WI, USA). Stained slides were ana-lyzed for the presence or absence of inflammation andfatty infiltration. All the slides were reviewed by a board-certified pathologist. All the images were scanned at × 20magnification using an Olympus BX51inverted micro-scope (Olympus Scientific Solutions, Waltham, MA, USA)with a scale bar of 200 μm. The average adipocyte size ininfra- and suprapatellar fat was calculated by selecting 15random adipocytes in the different swine groups in eachcategory using ImageJ software (National Institutes ofHealth [NIH], Bethesda, MD, USA). Chondrocyte count-ing was done using three high-power field images, and theaverage number of clustered chondrocytes to totalchondrocytes per high-power field in each group wascalculated. Areas of fatty infiltration were scanned, andfatty infiltration within the matrix of muscle, tendon, andligament was calculated by measuring the area of alladipocytes present in the matrix followed by calculatingthe percentage of infiltration area per field.

Histological analysis for proteoglycans, collagens, andmatrix metalloproteinasesTissue sections were stained with Alcian blue andSafranin-O Fast Green staining for analysis of the cartil-age mucosubstance content in VDDef, VDSuff, andVDSupp swine cartilage. The stained slides were scannedby using an Olympus microscope at × 40 magnification.The results of the stained slides were graded in a blindedmanner by a board-certified pathologist on a scale of0–5 where 0 =minimal blue staining, 1 = very weak bluestaining, 2 = weak blue staining, 3 =moderate blue stain-ing, 4 = strong blue staining, and 5 = very strong bluestaining for Alcian blue, and on a scale of 0–5 where0 = minimal orange-red staining, 1 = very weak orange-redstaining, 2 = weak orange-red staining, 3 =moderate or-ange-red staining, 4 = strong orange-red staining, and

5 = very strong orange-red staining for Safranin-OFast Green staining [40].Additionally, the tissue sections were examined by

modified Russell-Movat pentachrome staining to assessthe collagen and mucosubstance content of cartilage inVDDef, VDSuff, and VDSupp swine. The slides weredouble-checked by the pathologist in a blinded manner.The stained slides were scanned with an Olympusmicroscope at × 20 magnification and graded on a scaleof 0–5 for collagen (yellow staining) and mucinous sub-stance (blue-green staining), where 0 =minimal yellow/blue-green staining, 1 = very weak yellow/blue-green stain-ing, 2 = weak yellow/blue-green staining, 3 =moderateyellow/blue-green staining, 4 = strong yellow/blue-greenstaining, and 5 = very strong yellow/blue-green staining.The tissue sections were also subjected to immunohisto-

chemical (IHC) staining for aggrecans and proteoglycans(aggrecans, decorins, and versicans). The IHC staining wasconducted as per the standard protocol in our laboratory.Antigoat aggrecan (sc-25674; Santa Cruz Biotechnology,Dallas, TX, USA) for aggrecans in a 1:50 dilution and anti-mouse decorin (6D6-c; Developmental Studies HybridomaBank, Iowa City, IA, USA) and antimouse versican(12C5-c; Developmental Studies Hybridoma Bank) anti-bodies in a 1:100 dilution were used as the primary anti-bodies. A biotinylated secondary antibody was used, anddiaminobenzidine was used as a fluorochrome to developthe slides. The slides were counterstained with hematoxylinto stain the nuclei. The slides were blindly checked by thepathologist. The slides were scanned at × 40 magnificationwith an Olympus microscope, and the total number ofimmunopositive cells for aggrecans, decorins, and versi-cans were counted on three different images for all tissuesusing ImageJ software. Cell density per square millimeterwas calculated in the VDDef, VDSuff, and VDSupp groups.

Immunofluorescence studiesDeparaffinization, rehydration, and antigen retrieval wereperformed prior to immunostaining. Immunofluorescentstaining was done as per the standard protocol in ourlaboratory. As primary antibodies, we used rabbit anti-TREM-1 (PAA213Po01; Cloud-Clone Corp., Houston, TX,USA), rabbit anti-TREM-2 (bs-2723R; Bioss Inc., Woburn,MA, USA), and mouse anti-CD14 (ab182032; Abcam,Cambridge, MA, USA) for macrophages, mouse antiadipo-nectin (ab22554; Abcam), and rabbit antileptin (ab16227;Abcam) at 1:200 dilution, as well as mouse anti-neutrophilelastase (sc-53388; Santa Cruz Biotechnology) at 1:50dilution. Alexa Fluor 594-conjugated (red) and AlexaFluor 488-conjugated (green) secondary antibodies (LifeTechnologies, Carlsbad, CA, USA) at 1:500 dilution wereused. The slides were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) to stain nuclei. Negative controlswere run by using an isotype antibody for each

Rai et al. Arthritis Research & Therapy (2016) 18:203 Page 3 of 17

fluorochrome. Immunofluorescence microscopy was car-ried out with an Olympus BX51 inverted fluorescencemicroscope. The images were scanned at × 20 magnifica-tion and were reviewed by a board-certified pathologist.The mean fluorescence intensity (MFI) for TREM-1,TREM-2, adiponectin, and leptin was quantified by usingImageJ software. Two images from each swine in eachgroup were taken for the measurement of MFI. The num-ber of TREM-2 and CD-14 dual-positive cells was countedin five high-power fields for each tissue in all swine, andthe average number of TREM-2 and CD-14 dual-positivecells per square millimeter was calculated for each group.The analysis for the collagen types I, II, and III content

in the VDDef, VDSuff, and VDSupp swine was carriedout by immunofluorescent staining. Antirabbit collagen I(sc-30136; Santa Cruz Biotechnology), antigoat collagenII (sc-389924; Santa Cruz Biotechnology), and antirabbitcollagen III (sc-28888; Santa Cruz Biotechnology) anti-bodies in 1:50 dilution were used as primary antibodies.Alexa Fluor 594-conjugated (red) and Alexa Fluor 488-conjugated (green) secondary antibodies (Life Technologies)at 1:500 dilution were used. DAPI was used to stain thenuclei. The stained slides were blindly double-checked bythe pathologist. MFI for collagen II was calculated usingthree separate images for all swine in each group. ImageJsoftware was used to analyze the MFI. Average MFI wascalculated for the three different groups of swine.The analysis for the matrix metalloproteinases (MMPs)

in the VDDef, VDSuff, and VDsupp swine was also carriedout by immunofluorescent staining. Antigoat MMP-1(sc-21731; Santa Cruz Biotechnology), antimouse MMP-2(sc-53630; Santa Cruz Biotechnology), antimouse MMP-8(sc-101450; Santa Cruz Biotechnology), antigoat MMP-9(sc-6840; Santa Cruz Biotechnology), and antimouseMMP-13 (sc-101564; Santa Cruz Biotechnology) anti-bodies in a 1:50 dilution were used as primary antibodies.Alexa Fluor 594-conjugated (red) and Alexa Fluor 488-conjugated (green) secondary antibodies (Life Technolo-gies) at 1:500 dilution were used. DAPI was used to stainthe nuclei. The stained slides were blindly double-checkedby the pathologist. MFI for MMP-9 was calculated usingthe three separate images for all swine in each group. Ima-geJ software was used to analyze the MFI. Average MFIwas calculated for the three different groups of swine.

Analysis of the vitamin D status with TREM-2 MFI,macrophage density, chondrocyte density and clustering,and adipocyte areaTo correlate vitamin D status with various parameters ofinterest, correlation analysis was performed. The averagevalues for the macrophage density in the VDDef, VDSuff,and VDSupp groups were converted into logarithmvalues to fit them in the range of values for other param-eters (to fit them into a graph).

Statistical analysisResults are presented as mean ± SD for each parameter.The Wilcoxon two-samples test was used for statisticalanalysis of the significance of differences between theVDDef, VDSuff, and VDSupp groups. A p value <0.05was considered significant.

ResultsBiochemical parametersThe serum levels of vitamin D at the beginning of thestudy were <15 ng/ml in the VDDef group, 17.53 ±1.13 ng/ml in the VDSuff group, and 31.93 ± 3.04 ng/mlin the VDSupp group. The experimental diets werestarted when the swine were at the age of 6 months.After 1 year of feeding the swine the experimental diet,the mean serum 25(OH)D levels were 8.1 ± 1.13 ng/mlin the VDDef group, 24.01 ± 0.72 ng/ml in the VDSuffgroup, and 52.58 ± 5.68 ng/ml in the VDSupp group.The corresponding average levels of serum parathyroidhormone were 38.57 ± 5.75 ng/ml, 13.03 ± 3.23 ng/ml,and 8.52 ± 2.19 ng/ml, respectively; serum calcium,8.60 ± 0.78 ng/ml, 10.02 ± 0.22 ng/ml, and 9.7 ± 0.05 ng/ml,respectively; and serum cholesterol, 512.01 ± 56.9 ng/ml,551.75 ± 45.03 ng/ml, and 461 ± 43.08 ng/ml, respectively.The corresponding average weights at the time theanimals were killed were 95.96 ± 4.53 lb, 112.24 ± 6.23 lb,128.11 ± 13.97 lb, respectively (see Additional file 1).

Increased inflammation in vitamin D-deficient groupMicroscopic analysis of H&E-stained sections of infrapa-tellar fat (Fig. 1a, b, c), quadriceps muscle (Fig. 1d, e, f ),patellar tendon (Fig. 1j, k, l), collateral ligament(Fig. 1m, n, o), meniscal cartilage (Fig. 1p, q, r), suprapatel-lar fat (Fig. 1s, t, u), and synovial membrane (Fig. 1v, w, x)revealed increased inflammation in the VDDef groupcompared with the VDSuff and VDSupp groups. The in-flammation in the VDSuff group was higher than that inthe VDSupp group. In synovial membrane, there was mildinflammation and mild hyperplasia in the VDDef group,while the VDSuff and VDSupp groups showed minimalchanges. Inflammation was defined on the basis of thepresence of macrophages and/or neutrophils in the tissuematrix. Knee joint tissues in the VDDef group showed thepresence of many inflammatory cells along with the clus-ter of inflammation (>4 inflammatory cells), while therewere only occasional or few inflammatory cells in theVDSuff and VDSupp groups (see Additional file 2).

Increased fatty infiltration in vitamin D-deficient groupHistological analysis of H&E-stained sections of muscle,tendon, and ligament showed increased fatty infiltrationin the VDDef group (Fig. 2a, d, g) compared with theVDSuff (Fig. 2b, e, h) and VDSupp (Fig. 2c, f, i) groups.There was minimal fatty infiltration in the VDSuff

Rai et al. Arthritis Research & Therapy (2016) 18:203 Page 4 of 17

Fig. 1 Hematoxylin and eosin (H&E) staining of infrapatellar fat, quadriceps muscle, articular cartilage, patellar tendon, collateral ligament, meniscalcartilage, suprapatellar fat, and synovial membrane. H&E staining shows the histology and inflammation in the vitamin D-deficient (VDDef), vitaminD-sufficient (VDSuff), and vitamin D-supplemented (VDSupp) group infrapatellar fat (a, b, c), muscle (d, e, f), articular cartilage (g, h, i), tendon (j, k, l),ligament (m, n, o), menisci (p, q, r), suprapatellar fat (s, t, u), and synovial membrane (v, w, x). Arrows indicate the presence of inflammatory cells inthe tissue. These are the representative images of five swine each in the VDDef and VDSuff groups and three swine in the VDSupp group

Rai et al. Arthritis Research & Therapy (2016) 18:203 Page 5 of 17

group, and no fatty infiltration was observed in theVDSupp group (see Additional file 2). Fatty infiltrationwas defined as the presence of adipocytes betweenmuscle, tendon, and ligament fibers (i.e., not simply nextto the structure, but interspersed within the structure).The percentage of fatty infiltration per field wascalculated in the VDDef, VDSuff, and VDSupp groups(Fig. 2s). The presence of adipocytes interspersed withinthe structure was further confirmed by Oil Red O stain-ing and immunopositivity for adiponectin around theadipocyte (data not shown).

Increased chondrocyte clustering in vitamin D-deficientgroupHistological analysis of articular cartilage in the VDDef,VDSuff, and VDSupp groups revealed increased chondro-cyte clustering in the VDDef group compared with theVDSuff and VDSupp groups (Fig. 2j, k) as well as in theVDSuff group (Fig. 2k) compared with the VDSupp group(Fig. 2l). Quantitation of chondrocytes per high-powerfield showed that nearly 50 % of chondrocytes were

clustered in the VDDef and VDSuff groups, whereas only3–4 % were clustered in the VDSupp group (Figs. 2tand 3c) (see Additional file 1).

Increased adipocyte size in vitamin D-deficient groupAnalysis of the average adipocyte area in suprapatellar(Fig. 2m, n, o) and infrapatellar (Fig. 2p, q, r) fat pads re-vealed significantly increased adipocyte size in theVDDef group compared with the VDSuff and VDSuppgroups, as well as in the VDSuff group compared withthe VDSupp group (Fig. 2u).

Decreased expression of proteoglycans invitamin D-deficient groupAnalysis of Alcian blue, Safranin-O Fast Green, andmodified Russell-Movat pentachrome staining andimmunohistochemistry for proteoglycans (aggrecans,decorins, and versicans) showed decreased expression ofmatrix substance and proteoglycans in the VDDef group(Fig. 4a, d, g, j, m, p) compared with the VDSuff(Fig. 4b, f, h k, n and q) and VDSupp (Fig. 4c, f, i, l, o, r)

Fig. 2 Hematoxylin and eosin (H&E) staining of quadriceps muscle, patellar tendon, collateral ligaments, articular cartilage, and infra- and suprapatellar fatpads. H&E staining shows the fatty infiltration in the vitamin D-deficient (VDDef), vitamin D-sufficient (VDSuff), and vitamin D-supplemented (VDSupp)groups: quadriceps muscle (a, b, c), patellar tendon (d, e, f), and collateral ligaments (g, h, i). Chondrocyte clustering (×40 original magnification) in theVDDef and VDSuff groups compared with the VDSupp group (j, k, l), as well as the adipocyte area in infrapatellar (m, n, o) and suprapatellar fat pads(p, q, r), is shown. The percentage of fatty infiltration per high-power field (s), average total and clustered chondrocytes (t), and average adipocyte area(u) are graphed. Arrows indicate the presence of adipocytes within the matrix of the tissue and chondrocyte clustering. These are the representativeimages of five swine each in the VDDef and VDSuff groups and three swine in the VDSupp group. Data are shown as mean ± SD. *p < 0.05, **p< 0.01

Rai et al. Arthritis Research & Therapy (2016) 18:203 Page 6 of 17

groups. Comparison of the VDSuff group with theVDSupp group revealed increased expression of proteo-glycans in the VDSuff group compared with the VDSuppgroup (Fig. 4). Proteoglycan and matrix substance ex-pression in the VDSuff group swine was nonuniform,whereas the expression in the VDSupp group was uniform(Fig. 4, Table 1).

Increased leptin and decreased adiponectin expression invitamin D-deficient swineImmunofluorescence studies of infrapatellar and supra-patellar fat revealed greater immunoreactivity for leptinand lesser immunoreactivity for adiponectin in theVDDef group (Fig. 5a, j; Additional file 3a, j) than in the

VDSuff (Fig. 5d, m; Additional file 3d, m) and VDSupp(Fig. 5g, p; Additional file 3 g, p) groups. The MFI of lep-tin was higher and that of adiponectin was lower in theVDDef group than in the VDSuff and VDSupp groups(Fig. 5s, t). The average adipocyte counts in infra- andsuprapatellar fat pads were higher in the VDDef groupthan in the VDSuff and VDSupp groups, and they werehigher in the VDSuff group than in the VDSupp group(Fig. 5s, t).

Increased expression of TREM-2 in vitamin D-sufficientand vitamin D-supplemented groupsImmunofluorescence studies of articular cartilage(Fig. 6a, e, i), synovial membrane (Fig. 6a, q, u),

Fig. 3 Analysis of changes in inflammation, fatty infiltration, adipocyte area, and chondrocyte clustering by vitamin D status in knee joint tissues.a Macrophage density (per square millimeter, logarithm values), triggering receptor expressed on myeloid cells (TREM)-2 mean fluorescenceintensity (MFI), and adipocyte area (per square millimeter) in infrapatellar fat. b Macrophage density (per square millimeter), TREM-2 MFI, and fattyinfiltration (percentage of area, logarithm values) in muscle. c Macrophage density (per square millimeter, logarithm values), TREM-2 MFI, and chondrocyteclustering (clustered/total chondrocytes, logarithm values) in articular cartilage. d Macrophage density (per square millimeter), TREM-2 MFI, and fattyinfiltration (percentage of area, logarithm values) in tendon. e Macrophage density (per square millimeter), TREM-2 MFI, and fatty infiltration (percentageof area) in ligament. f Macrophage density (per square millimeter) and TREM-2 MFI in menisci. g Macrophage density (per square millimeter, logarithmvalues), TREM-2 MFI, and adipocyte area (per square millimeter) in suprapatellar fat. h Macrophage density (per square millimeter, logarithm values) andTREM-2 MFI in synovial membrane

Rai et al. Arthritis Research & Therapy (2016) 18:203 Page 7 of 17

infrapatellar fat (see Additional file 4a, e, i), quadricepsmuscle (see Additional file 4m, q, u), patellar tendon (seeAdditional file 5a, e, i), collateral ligaments (see Additionalfile 5m, q, u), meniscal cartilage (see Additional file 6a, e, i),and suprapatellar fat (see Additional file 6 m, q, u) re-vealed higher immunoreactivity for TREM-2 in theVDSupp group than in the VDSuff and VDDef groups andin the VDSuff group than in the VDDef group (Fig. 3).The MFI of TREM-2 was significantly higher (Fig. 7a) inthe VDSuff and VDSupp groups than in the VDDef group.The colocalization of TREM-2 with macrophages was ob-served more in the VDDef group than in the VDSuff and

VDSupp groups and in the VDSuff group than in theVDSupp group. We found very minimal positivity forTREM-1 in articular cartilage, tendon, ligament, andsuprapatellar fat (see Additional file 7a, e, i, m).

Decreased expression of collagen II and increasedexpression of MMP-9 in vitamin D-deficient groupImmunofluorescent staining for collagen I, II, and III re-vealed higher immunoreactivity of collagen II than forcollagen I and III in all cartilage tissues (see Additionalfile 8a, d, g). The immunofluorescence studies forcollagen II revealed decreased immunoreactivity in the

Fig. 4 Alcian blue, Safranin-O Fast Green, modified Russell-Movat pentachrome, and immunohistochemical staining for matrix substance andproteoglycans (aggrecans, decorins, and versicans). Alcian blue (a, b, c), Safranin-O Fast Green (d, e, f), modified Russel-Movat pentachrome(g, h, i), aggrecans (j, k, l), decorins (m, n, o), and versicans (p, q, r) in the vitamin D-deficient (VDDef), vitamin D-sufficient (VDSuff), and vitaminD-supplemented (VDSupp) groups of swine. These are the representative images of five swine each in the VDDef and VDSuff groups and threeswine in the VDSupp group

Rai et al. Arthritis Research & Therapy (2016) 18:203 Page 8 of 17

VDDef group compared with the VDSuff and VDSuppgroups (Fig. 8a, d, g), and in the VDSuff group comparedwith the VDSupp group (Fig. 8d, g). The immunofluores-cence studies showed no immunoreactivity for MMP-1,MMP-2, and MMP-13 (see Additional file 9a, d, m), veryminimal immunoreactivity for MMP-8 (see Additionalfile 9g), and higher immunoreactivity for MMP-9 (seeAdditional file 9j) in the VDDef, VDSuff, and VDSuppgroups. The immunoreactivity of MMP-9 was higherin the VDDef group than in the VDSuff and VDSuppgroups (Fig. 8j, m, p) and in the VDSuff group than in theVDSupp group (Fig. 8m, p). The MFI of collagen II(Fig. 8s) and MMP-9 (Fig. 8t) was significantly higher inthe VDDef group than in the VDSuff and VDSupp groupsand in the VDSuff group than in the VDSupp group.

Increased number of macrophages in vitamin D-deficientgroupDual-immunofluorescence studies of TREM-2 andmacrophage for suprapatellar and infrapatellar fat,muscle, tendon, ligament, meniscal cartilage, articularcartilage, and synovial membrane showed significantlyan increased number of macrophages in the VDDefgroup compared with the VDSuff and VDSupp groups(Fig. 3). Higher numbers of CD14+ cells were foundcolocalized with TREM-2 in the VDDef group than inthe VDSuff and VDSupp groups (Fig. 7b). Fewer

macrophages colocalized with TREM-1 in articular car-tilage, tendon, ligament, and suprapatellar fat in theVDDef group were also observed (see Additional file 7).Dual-immunofluorescence of TREM-1, TREM-2, andneutrophils showed no positivity for neutrophil elastase,suggesting the absence of neutrophils in the knee jointtissues (data not shown).

DiscussionOn the basis of microscopic analysis of H&E staining, com-pared with the VDSupp group, in the VDDef and VDSuffgroups we found increased inflammation in knee jointtissues; increased fatty infiltration in quadriceps muscle,patellar tendon, and collateral ligaments; and increasedchondrocyte clustering. These findings were further sup-ported by immunofluorescence studies. Histological ana-lysis of the matrix substance, proteoglycans, and collagensshowed increased cartilage matrix loss in the VDDef groupcompared with the VDSuff and VDSupp groups and in theVDSuff group compared with the VDSupp group. Therewas increased expression of MMP-9 in the VDDef groupcompared with the VDSuff and VDSupp groups and in theVDSuff group compared with the VDSupp group. The veryweak light blue staining in VDDef group compared withVDSuff and VDSupp group, as well as the moderate bluestaining in the VDSuff group compared with very strongblue staining in the VDSupp group by Alcian blue staining,

Table 1 Histological analysis of matrix substance and proteoglycans (aggrecan, decorin, and versican) for collagen loss in kneecartilage of VDDef, VDSuff, and VDSupp swine

Swine group Alcian blue staining Safranin-O and FastGreen staining

Modified Russell-Movatpentachrome staining(blue-green/yellow)

Aggrecan+ cells/mm2 Decorin+ cells/mm2 Versican+ cells/mm2

VDDef

1 3 0 3/4 4.72 ± 0.84 12.16 ± 1.39 5.06 ± 1.23

2 1 0 2/4 0 0 0

3 2 0 3/3 0 0 0

4 2 2 3/4 18.51 ± 4.42 11.29 ± 0.41 10.4 ± 1.25

5 2 3 3/4 20.58 ± 2.85 11.16 ± 1.69 8.34 ± 0.52

VDSuff

1 5 2 2/4 31.95 ± 4.23 56.22 ± 2.88 28.11 ± 1.90

2 2 2 4/2 20.18 ± 2.88 17.54 ± 3.01 13.38 ± 1.51

3 2 3 3/3 18.5 ± 1.5 18.97 ± 3.56 17.98 ± 1.20

4 2 5 3/1 51.18 ± 2.59 19.46 ± 3.3 20.90 ± 3.14

5 5 3 4/1 63.79 ± 8.19 29.85 ± 2.27 15.37 ± 1.5

VDSupp

1 4 3 5/1 36.95 ± 3.48 15.78 ± 0.67 19.26 ± 1.69

2 4 4 5/1 27.54 ± 1.87 10.57 ± 2.2 9.61 ± 2.31

3 4 4 5/2 28.74 ± 1.88 11.05 ± 1.1 8.41 ± 1.81

Abbreviations: VDDef Vitamin D-deficient, VDSuff Vitamin D-sufficient, VDSupp Vitamin D-supplementedAlcian blue, Safranin-O Fast Green, and modified Russell-Movat pentachrome staining were done for assessment of cartilage mucosubstance damage.Immunohistochemical staining was done for aggrecans, decorins, and versicans, and the immunopositive cell density was calculated. Data are presented asmean ± SD for immunopositive cell density. See the Methods section for discussion of cartilage grading

Rai et al. Arthritis Research & Therapy (2016) 18:203 Page 9 of 17

suggests loss of sulfated mucosubstances of the cartilagewith decreasing vitamin D. Similarly, absence of orange tored coloring in the VDDef group and presence of orange-red coloring in VDSuff and VDSupp group suggest loss ofcartilage and mucinous matrix substance. In modifiedRussell-Movat pentachrome staining, the VDDef groupshowed more yellow coloration than the VDSuff andVDSupp groups, while the VDSuff group showed blue-green coloration and the VDSupp group had strong bluecoloration. The predominance of yellow coloration inVDDef group may be due to the loss of matrix mucosub-stance and exposure of the collagen fibrils, whereas in theVDSuff and VDSupp groups, strong blue coloration in-dicating the presence of matrix mucosubstance mayhave masked the yellow coloration of the collagen fibers,

as they were not exposed. These results suggest thebeneficial role of vitamin D supplementation in protect-ing the cartilage matrix by increasing matrix substanceand proteoglycans.Inflammation and fatty infiltration may render the

muscle weak and contribute to OA [2, 6]. The architec-tural changes or distortion of the normal architecture inquadriceps muscles, patellar tendon, and collateral liga-ments with the foci of inflammation and fatty infiltrationsuggest that inflammation and fatty infiltration canchange the knee joint morphology in vitamin D defi-ciency. Absence of architectural changes and fatty infil-tration in the VDSupp group suggests the potentialbeneficial role of vitamin D in decreasing inflammationand fatty infiltration in knee joint tissues. These results

Fig. 5 Immunofluorescent staining for adiponectin and leptin in infra- and suprapatellar fat pads. Immunofluorescent staining in the vitaminD-deficient (VDDef), vitamin D-sufficient (VDSuff), and vitamin D-supplemented (VDSupp) group infrapatellar fat is shown for adiponectin (a, d, g),leptin (j, m, p), 4′,6-diamidino-2-phenylindole (DAPI) (b, e, h, k, n, q), merged adiponectin-DAPI (c, f, i), and merged leptin-DAPI (l, o, r). The meanfluorescence intensity (MFI) of adiponectin and leptin with average adipocyte count (s) in infrapatellar fat, and the MFI of adiponectin and leptinwith average adipocyte count in suprapatellar fat (t), is shown for VDDef, VDSuff, and VDSupp swine. These are the representative images of fiveswine each in the VDDef and VDSuff groups and three swine in the VDSupp group. Data are shown as mean ± SD. *p < 0.05, **p < 0.01

Rai et al. Arthritis Research & Therapy (2016) 18:203 Page 10 of 17

are in accordance with those of previous studies suggest-ing a restorative effect on muscle weakness with vitaminD supplementation [6]. Distortion of muscle structuremay lead to muscle weakness, and it has been suggestedthat muscle weakness is an early symptom of OA. Lowlevels of vitamin D are associated with muscle weakness

and impaired function [41]. Muscle weakness results indecreased shock-absorbing capacity in knee joint andprecedes and mediates cartilage degeneration [2, 42].Our findings suggest a role for vitamin D supplementa-tion in decreasing inflammation in tendon and ligamentin addition to its beneficial effects on muscle.

Fig. 6 Immunofluorescent staining for triggering receptor expressed on myeloid cells (TREM)-2 and macrophage (CD14) as well as colocalization ofTREM-2 with CD14+ cells in articular cartilage and synovial membrane. Immunofluorescent staining for TREM-2 (a, e, i, m, q, u), CD14 (b, f, j, n, r, v),4′,6-diamidino-2-phenylindole (DAPI) (c, g, k, o, s, w), and merged images for examining colocalization of TREM-2 and CD14 (d, h, l, p, t, x). Arrowsindicate colocalization of TREM-2 with CD14. These are the representative images of five swine each in the vitamin D-deficient (VDDef) and vitaminD-sufficient (VDSuff) groups and three swine in the vitamin D-supplemented (VDSupp) group

Rai et al. Arthritis Research & Therapy (2016) 18:203 Page 11 of 17

Obesity has long been considered a risk factor for OA.Due to the extension of the metabolic effects of obesityto the fat depots in the knee joint, suprapatellar fat(SPF), and infrapatellar fat (IPF), inflammation in SPFand IPF may play a significant role in knee joint or patel-lofemoral joint OA [43, 44]. White adipocytes acting asendocrine cells secrete adipokines and cytokines thatregulate various functions in the human body, includingangiogenesis and the immune response [45]. Histologicalanalysis of knee joint tissues revealed inflammation inIPF and SPF, mainly in VDDef swine compared withminimal inflammation in the VDSuff group and no in-flammation in the VDSupp group. We further stainedthese tissues for the adipokines adiponectin and leptin.Immunofluorescence studies revealed lower levels of adi-ponectin and higher levels of leptin in the VDDef groupthan in the VDSuff and VDSupp groups. Adiponectinmay play a protective role in the pathogenesis of OA byincreasing expression of tissue inhibitors of metallopro-teinases and decreasing MMPs, and it may possess sev-eral anti-inflammatory properties by suppressing nuclearfactor kB, interleukin 6, and tumor necrosis factor α [7].

Adiponectin levels have been considered as a biomarkerassociated with early radiographic disease progression inearly rheumatoid arthritis (RA) [46]. Further, increasedadiponectin levels in synovial fluid are associated withRA and are thought to counteract the local inflamma-tory process [47]. Increased expression of adiponectin inthe VDSupp group in this study may also suggest a roleof vitamin D in decreasing inflammation and increasingproduction of adiponectin to protect the knee joint. Fur-ther, the role of vitamin D deficiency in other forms of in-flammatory arthritis, such as RA, has been discussed, andvitamin D intake has been negatively correlated with thedevelopment of RA [48]. Increased levels of leptin are det-rimental to chondrocytes and are associated with cartilagedegeneration and volume loss [31, 32]. In accordance withthese studies, we also found decreased adiponectin and in-creased leptin levels in the VDDef group compared withthe VDSuff and VDSupp groups, suggesting a role of vita-min D in enhancing adiponectin and attenuating leptinlevels, probably by decreasing inflammation.Fatty infiltration of quadriceps muscle, patellar tendon,

and collateral ligaments was associated with increased

Fig. 7 Analysis of triggering receptor expressed on myeloid cells (TREM)-2 immunoreactivity and macrophage density. Mean fluorescenceintensity of TREM-2 (a) and number of macrophages (CD14+) (b) in the vitamin D-deficient (VDDef) (n = 5), vitamin D-sufficient (VDSuff) (n = 5),and vitamin D-supplemented (VDSupp) (n = 3) group infrapatellar fat, muscle, articular cartilage, tendon, ligament, menisci, suprapatellar fat, andsynovial membrane. A comparative analysis was performed between the VDDef, VDSuff, and VDSupp groups. Data are shown as mean ± SD.*p < 0.05, **p < 0.01

Rai et al. Arthritis Research & Therapy (2016) 18:203 Page 12 of 17

inflammation in areas of fatty infiltration in the VDDefand VDSuff groups. There was only minimal inflamma-tion in the VDSupp group. Immunofluorescence studiesof knee joint tissues for TREM-1 and TREM-2 showedminimal immunoreactivity for TREM-1, but, to our sur-prise, all the tissues in the VDDef, VDSuff, and VDSuppgroups showed immunoreactivity for TREM-2, althoughTREM-2 immunoreactivity was higher in the VDSuppgroup than the VDDef and VDSuff groups. In agreementwith the literature which suggests that TREM-1 expres-sion increases during inflammation [27–29], in our studyonly few tissues showed TREM-1 immunoreactivity. Thehigh TREM-2 immunoreactivity in the VDDef group in

inflamed tissues was of interest, so we performed a dual-immunofluorescence study of TREM-1 and TREM-2 withmacrophages and neutrophils. Dual-immunofluorescencerevealed higher number and more colocalization of mac-rophages with TREM-2 in the VDDef group. This suggeststhat higher TREM-2 expression in the presence of macro-phages may be due to the protective response or to innateimmunity of the body. Because macrophage populationsare highly heterogeneous, changes in heterogeneity ofmacrophages may play a role in OA. The lower immuno-reactivity for TREM-1 and higher immunoreactivity forTREM-2 also suggest a predominance of M2 macro-phages, the anti-inflammatory macrophages, compared

Fig. 8 Immunofluorescence studies for collagen type II and matrix metalloproteinase 9 (MMP-9) in cartilage. Immunofluorescent staining isshown for collagen II (a, d, g), MMP-9 (j, m, p), 4′,6-diamidino-2-phenylindole (DAPI) (b, e, h, k, n, q), and merged images (c, f, i, l, o, r). Meanfluorescence intensity is shown for collagen II (s) and MMP-9 (t). These are the representative images of five swine each in the vitamin D-deficient(VDDef) and vitamin D-sufficient (VDSuff) groups and three swine in the vitamin D-supplemented (VDSupp) group. Data are shown as mean ± SD.**p < 0.01, ***p < 0.001

Rai et al. Arthritis Research & Therapy (2016) 18:203 Page 13 of 17

with M1 macrophages, which are proinflammatory. Thechange in heterogeneity of the macrophage plays a role inbone remodeling or osteogenesis by changing the micro-environment [49], but the role of heterogeneity in thepathogenesis of OA needs to be determined. Further, theabsence of neutrophils and the presence of macrophagesonly as the inflammatory cells suggest the chronicity of in-flammation in the knee joint [50, 51] and a possible roleof macrophages in the inflammatory response. Vitamin Dmay have a pleiotropic effect in immune cells, includingmacrophages, and an immunomodulatory role of vitaminD in various immune cells and diseases has been discussed[52]. The role of vitamin D in blocking the inflammatorycytokines, cell proliferation, and angiogenesis in the con-text of various diseases has been studied, and potentialbeneficial immunomodulatory and anti-inflammatoryroles have been elucidated [53, 54]. In this study, thehigher expression of TREM-2 in the VDSupp group sug-gests an anti-inflammatory role of vitamin D in knee jointtissues. The decreased colocalization of TREM-2 withmacrophages in dual-immunofluorescence studies in theVDSupp group compared with the VDDef group furthersupports the anti-inflammatory role of vitamin D. Thesefindings are in accordance with previous studies sug-gesting decreased vitamin D receptor expression duringinflammation and an inflammation-attenuating effect ofvitamin D supplementation [33, 55].Loss of cartilage volume and subchondral bone is the

primary pathophysiology of OA. Prevention of bonedamage to reduce progression of cartilage damage andloss may reduce the severity and progression of OA [56],and vitamin D can play an important role in attenuatingbone damage [14]. Clustering of chondrocytes in degen-erating cartilage has been suggested as a hallmark of OAand classified as grade 2 cartilage degeneration by theInternational Cartilage Repair Society [57, 58]. Chondro-cyte clustering is also a feature of autoimmune arthritis(i.e., RA), where chronic inflammation does play a rolein the pathogenesis. Clustering of cells has also beendiscussed in other degenerative joint diseases involvingintervertebral discs, menisci, and cricoarytenoid cartil-age. Chondrocyte clustering may be due to the diseaseprogression in OA or to the increased proliferation ormigration of cells from the deeper layers [59]. In thisstudy, we found chondrocyte clustering (50 % of the totalchondrocytes) in the VDDef and VDSuff groups and only3 % clustering in the VDSupp group. We also observedthat chondrocyte clustering was more prominent towardthe edges of cartilage. These data suggest that vitamin Dsupplementation may have led to the decreased clusteringin the VDSupp group. As the chondrocyte clustering isassociated with cartilage degeneration, vitamin D supple-ments may act as an intervening agent to decrease theprogression of cartilage loss. It has also been suggested

that clustered chondrocytes near degenerating cartilagehave progenitor characteristics with proliferative potential[58]; vitamin D as a proliferative agent may thus act as anadjunctive therapy.Matrix degeneration occurs as a result of increased

levels of degrading enzymes, and the activity of thesemetalloproteinase enzymes is modulated by vitamin D.Low levels of vitamin D may lead to increased produc-tion of MMPs [60]. In this study, we found increasedlevels of MMP-9 expression in VDDef swine comparedwith VDSuff and VDSupp swine. Further, the MMP-9levels were higher in VDSuff swine than in VDSuppswine. These results suggest that vitamin D supplemen-tation attenuated the expression of MMP-9 in cartilage.The expression of collagen II was decreased in theVDDef group compared with the VDSuff and VDSuppgroups and in the VDSuff group compared with theVDSupp group, which can be explained by the higherlevels of MMP-9 in VDDef and VDSuff swine. Further,vitamin D deficiency is associated with smaller proteogly-can monomers [61], and our results showing decreasedproteoglycans (aggrecans, decorins, and versicans) inVDDef swine support the loss of proteoglycans in thatgroup. The increased expression of proteoglycans in theVDSuff and VDSupp groups suggests a beneficial role ofvitamin D. There was uniformity of expression of matrixsubstance and proteoglycans in the VDSupp group, butthe expression was nonuniform in VDSuff swine. Thehigher expression of matrix substance and proteoglycansin two swine in the VDSuff group can be explained by de-creased aggrecan expression through its downregulationat the posttranscriptional level with vitamin D3 [62].Vitamin D3 may also potentiate the MMP expression inactivated chondrocytes, facilitating intrinsic chondrolysisin the presence of proinflammatory cytokines, and it maysuppress the MMPs in activated synovial fibroblasts to de-crease chondrolysis. Hence, vitamin D may not directlyaffect the MMPs and can modulate production of othercytokines involved in the process of chondrolysis [63].Although serum vitamin D at baseline levels has an asso-ciation with OA, a low vitamin D level did not predict theloss of articular cartilage or narrowing of the joint space[64]. However, a higher level of vitamin D has beenassociated with a higher rate of hospitalization forknee or hip joint OA [65]. Hence, there is a need tofurther research the role of vitamin D in the patho-genesis of OA.

Limitations of the studyA limitation of this study was the absence of control do-mestic swine fed a normal diet. Another limitation maybe the small sample size in the VDSupp group. Despitethese limitations, the attenuation of inflammation andfatty infiltration in the VDSupp group by vitamin D

Rai et al. Arthritis Research & Therapy (2016) 18:203 Page 14 of 17

provides a new basis for further research into the role ofvitamin D in OA pathogenesis.

ConclusionsVitamin D supplementation is associated with attenuationof inflammation, fatty infiltration, chondrocyte clustering,and preservation of the tissue architecture. The histo-logical and immunofluorescence results of the presentstudy are closely related and thus validate the microscopicfindings. The results of this study demonstrate the poten-tial beneficial effect of vitamin D in decreasing inflamma-tion and fatty infiltration in knee joints, which maydecrease pain and disability. However, the role of vitaminD supplementation in decreasing the progression of cartil-age loss or OA needs to be investigated.

Additional files

Additional file 1: Table S2. Baseline table for swine groups. The studyincluded a total of 13 swine: 5 in the VDDef group, 5 in the VDSuffgroup, and 3 in the VDSupp group. Weight is indicated in pounds (lb).Average vitamin D3 levels were measured after 1 year of vitamin D dietintervention. Chondrocyte clustering was measured at × 40 magnification.(DOCX 13 kb)

Additional file 2: Table S3. Overview of all tissues with regard toinflammation and fatty infiltration. Inflammation in knee joint tissues(suprapatellar fat, infrapatellar fat, muscle, tendon, ligament, and menisci)was subjectively graded on the basis of the following criteria: nil = noinflammation; 1+ = occasional inflammatory cells; 2+ = few inflammatorycells; 3+ =many inflammatory cells; and 4+ = clumps or clusters ofinflammatory cells. Macrophage density was calculated for inflammationin cartilage of all knee joints. Macrophage density (CD14+ cells) in all theknee joint tissues was analyzed for objective grading of inflammation(Fig. 6). Subjective grading of fatty infiltration was based on the followingcriteria: no fatty infiltration; minimal fatty infiltration = fatty infiltration inup to 5 % of tissue area; mild fatty infiltration = fatty infiltration in 6–25 %of tissue area; moderate fatty infiltration = fatty infiltration in 25–50 % oftissue area; and severe fatty infiltration = fatty infiltration in >50 % oftissue area. Further, the percentage of fatty infiltration was calculated(Fig. 2). (DOCX 16 kb)

Additional file 3: Figure S9. Immunofluorescent staining for adiponectinand leptin in suprapatellar fat pad. Immunofluorescent staining in the VDDef,VDSuff, and VDSupp suprapatellar fat pads for adiponectin (a, d, and g), leptin(j,m, and p), DAPI (b, e, h, k, n, and q), merged adiponectin-DAPI (c, f, and i),and merged leptin-DAPI (l, o, and r). These are the representative images offive swine each in the VDDef and VDSuff groups and three swine in theVDSupp group. (TIF 91287 kb)

Additional file 4: Figure S10. Immunofluorescent staining for TREM-2and macrophage (CD14) and colocalization of TREM-2 with CD14+ cellsin infrapatellar fat and muscle. Immunofluorescent staining for TREM-2(a, e, i, m, q, and u), CD14 (b, f, j, n, r, and v), and DAPI (c, g, k, o, s,and w) was performed, and stains were merged to examine colocalizationof TREM-2 and CD14 (d, h, l, p, t, and x). Arrows show the colocalization ofTREM-2 with CD14. These are the representative images of five swine eachin the VDDef and VDSuff groups and three swine in the VDSupp group.(TIF 159298 kb)

Additional file 5: Figure S11. Immunofluorescent staining for TREM-2and macrophage (CD14) as well as colocalization of TREM-2 with CD14+

cells in tendon and ligament. Immunofluorescent staining for TREM-2(a, e, i, m, q, and u), CD14 (b, f, j, n, r, and v), and DAPI (c, g, k, o, s,and w) was performed, and stains were merged to examine colocalizationof TREM-2 and CD14 (d, h, l, p, t, and x). Arrows show the colocalization ofTREM-2 with CD14. These are the representative images of five swine each

in the VDDef and VDSuff groups and three swine in the VDSupp group.(TIF 157582 kb)

Additional file 6: Figure S12. Immunofluorescent staining for TREM-2and macrophage (CD14) and for colocalization of TREM-2 with CD14+ cellsin menisci and suprapatellar fat. Immunofluorescent staining for TREM-2(a, e, i, m, q, and u), CD14 (b, f, j, n, r, v), and DAPI (c, g, k, o, s, and w) wasperformed, and stains were merged to examine colocalization of TREM-2and CD14 (d, h, l, p, t, and x). Arrows show the colocalization of TREM-2 withCD14. These are the representative images of five swine each in the VDDefand VDSuff groups and three swine in the VDSupp group. (TIF 157634 kb)

Additional file 7: Figure S13. Immunofluorescent staining for TREM-1and macrophage (CD14) and for colocalization of TREM-1 with CD14+ cells inarticular cartilage, tendon, ligament, and suprapatellar fat. Immunofluorescentstaining for TREM-1 (a, e, i, and m), CD14 (b, f, j, and n), and DAPI (c, g, k,and o) was performed, and stains were merged to examine colocalizationof TREM-1 and CD14 (d, h, l, and p). Arrows show the colocalization ofTREM-1 with CD14. These are the representative images of five swine eachin the VDDef and VDSuff groups and three swine in the VDSupp group.(TIF 107810 kb)

Additional file 8: Figure S14. Immunofluorescent staining for collagentypes I, II, and III in articular cartilage. Immunofluorescent staining forcollagen I (a), collagen II (d), collagen III (g), and DAPI (b, e, and h).Merged images are also shown (c, f, and i). These are the representativeimages of five swine each in the VDDef and VDSuff groups and threeswine in the VDSupp group. (TIF 79516 kb)

Additional file 9: Figure S15. Immunofluorescent staining for MMPs 1,2, 8, 9, and 13 in articular cartilage. Immunofluorescent staining for MMP-1(a), MMP-2 (d), MMP-8 (g), MMP-9 (j), MMP-13 (m), and DAPI (b, f, h, k, n,and q), as well as merged images (c, f, i, l, o, and r), are shown. These arethe representative images of five swine each in the VDDef and VDSuffgroups and three swine in the VDSupp group. (TIF 75620 kb)

AbbreviationsDAPI: 4′,6-Diamidino-2-phenylindole; H&E: Hematoxylin and eosin;IHC: Immunohistochemical; IPF: Infrapatellar fat; MFI: Mean fluorescenceintensity; MMP: Matrix metalloproteinase; OA: Osteoarthritis; RA: Rheumatoidarthritis; SPF: Suprapatellar fat; TREM: Triggering receptor expressed onmyeloid cells; VDDef: Vitamin D-deficient; VDR: Vitamin D receptor;VDSuff: Vitamin D-sufficient; VDSupp: Vitamin D-supplemented;

AcknowledgementsWe acknowledge Ashley R. Pike for providing technical help during this study.

FundingThis study was supported by National Institutes of Health grants R01 HL116042,R01 HL112597, and R01 HL120659 (to DKA).

Availability of data and materialsIncluded in the paper.

Authors’ contributionsVR and DKA conceived of and designed the study. DKA contributedreagents, materials, and analytical tools. VR, NED, and MFD analyzed andinterpreted the data. VR, NED, MFD, MMR, and DKA drafted the manuscriptand revised it critically for important intellectual content. All authors readand approved the final manuscript.

Competing interestsThe corresponding author declares that this article is original; that the articledoes not infringe upon any copyright or other proprietary right of any thirdparty; and that neither the text nor the data have been reported or publishedpreviously. All of the authors declare that they have no competing interests andhave read the journal’s authorship statement.

Consent for publicationNot applicable.

Rai et al. Arthritis Research & Therapy (2016) 18:203 Page 15 of 17

Ethics approval and consent to participateThe research protocol (IACUC 0971) for conducting the study was approvedby the Institutional Animal Care and Use Committee of Creighton University.Yucatan microswine (30–40 lb) were purchased from Sinclair Laboratories,Columbia, MO, USA. Swine were kept in a controlled environmentthroughout the study in the Animal Resource Facility of Creighton University,Omaha, NE, USA. NIH standards and U.S. Department of Agricultureguidelines were followed for their care and experimental protocol.

Author details1Department of Clinical and Translational Science, Creighton UniversitySchool of Medicine, CRISS II Room 510, 2500 California Plaza, Omaha, NE68178, USA. 2Department of Pathology, Creighton University School ofMedicine, 601 North 30th Street, Omaha, NE 68131, USA. 3Department ofOrthopedic Surgery, Creighton University School of Medicine, Omaha, NE68178, USA. 4CHI Health Alegent Creighton Clinic, 601 North 30th Street,Suite 2300, Omaha, NE 68131, USA.

Received: 11 May 2016 Accepted: 22 August 2016

References1. Zhang Y, Jordan JM. Epidemiology of osteoarthritis. Clin Geriatr Med.

2010;26(3):355–69.2. Barker T, Henriksen VT, Rogers VE, Aguirre D, Trawick RH, Lynn Rasmussen G,

et al. Vitamin D deficiency associates with γ-tocopherol and quadricepsweakness but not inflammatory cytokines in subjects with kneeosteoarthritis. Redox Biol. 2014;2:466–74.

3. Cao Y, Winzenberg T, Nguo K, Lin J, Jones G, Ding C. Association betweenserum levels of 25-hydroxyvitamin D and osteoarthritis: a systematic review.Rheumatology (Oxford). 2013;52(7):1323–34.

4. Kwan Tat S, Lajeunesse D, Pelletier JP, Martel-Pelletier J. Targeting subchondralbone for treating osteoarthritis: what is the evidence? Best Pract Res ClinRheumatol. 2010;24(1):51–70.

5. Radin EL, Rose RM. Role of subchondral bone in the initiation andprogression of cartilage damage. Clin Orthop Relat Res. 1986;213:34–40.

6. Heidari B, Javadian Y, Babaei M, Yousef-Ghahari B. Restorative effect ofvitamin D deficiency on knee pain and quadriceps muscle strength in kneeosteoarthritis. Acta Med Iran. 2015;53(8):466–70.

7. Chen TH, Chen L, Hsieh MS, Chang CP, Chou DT, Tsai SH. Evidence for aprotective role for adiponectin in osteoarthritis. Biochim Biophys Acta.2006;1762(8):711–8.

8. Weinans H, Siebelt M, Agricola R, Botter SM, Piscaer TM, Waarsing JH.Pathophysiology of peri-articular bone changes in osteoarthritis. Bone.2012;51(2):190–6.

9. Tetlow LC, Woolley DE. Expression of vitamin D receptors and matrixmetalloproteinases in osteoarthritic cartilage and human articularchondrocytes in vitro. Osteoarthritis Cartilage. 2001;9(5):423–31.

10. Colombini A, Cauci S, Lombardi G, Lanteri P, Croiset S, Brayda-Bruno M,et al. Relationship between vitamin D receptor gene (VDR) polymorphisms,vitamin D status, osteoarthritis and intervertebral disc degeneration.J Steroid Biochem Mol Biol. 2013;138:24–40.

11. Holick MF, Chen TC. Vitamin D deficiency: a worldwide problem with healthconsequences. Am J Clin Nutr. 2008;87(4):1080S–6S.

12. Goula T, Kouskoukis A, Drosos G, Tselepis AS, Ververidis A, Valkanis C, et al.Vitamin D status in patients with knee or hip osteoarthritis in aMediterranean country. J Orthop Traumatol. 2015;16(1):35–9.

13. Al-Jarallah KF, Shehab D, Al-Awadhi A, Nahar I, Haider MZ, Moussa MA. Are25(OH)D levels related to the severity of knee osteoarthritis and function?Med Princ Pract. 2012;21(1):74–8.

14. Heidari B, Heidari P, Hajian-Tilaki K. Association between serum vitamin Ddeficiency and knee osteoarthritis. Int Orthop. 2011;35(11):1627–31.

15. Zhang FF, Driban JB, Lo GH, Price LL, Booth S, Eaton CB, et al. Vitamin Ddeficiency is associated with progression of knee osteoarthritis. J Nutr.2014;144(12):2002–8.

16. Ding C, Cicuttini F, Parameswaran V, Burgess J, Quinn S, Jones G. Serumlevels of vitamin D, sunlight exposure, and knee cartilage loss in olderadults: the Tasmanian older adult cohort study. Arthritis Rheum.2009;60(5):1381–9.

17. Bergink AP, Uitterlinden AG, Van Leeuwen JP, Buurman CJ, Hofman A,Verhaar JA, et al. Vitamin D status, bone mineral density, and the

development of radiographic osteoarthritis of the knee: the RotterdamStudy. J Clin Rheumatol. 2009;15(5):230–7.

18. Konstari S, Paananen M, Heliovaara M, Knekt P, Marniemi J, Impivaara O, et al.Association of 25-hydroxyvitamin D with the incidence of knee and hiposteoarthritis: a 22-year follow-up study. Scand J Rheumatol. 2012;41(2):124–31.

19. Muraki S, Dennison E, Jameson K, Boucher BJ, Akune T, Yoshimura N, et al.Association of vitamin D status with knee pain and radiographic kneeosteoarthritis. Osteoarthritis Cartilage. 2011;19(11):1301–6.

20. Wu S, Sun J. Vitamin D, vitamin D receptor, and macroautophagy ininflammation and infection. Discov Med. 2011;11(59):325–35.

21. Miller RR, Hicks GE, Shardell MD, Cappola AR, Hawkes WG, Yu-Yahiro JA,et al. Association of serum vitamin D levels with inflammatory responsefollowing hip fracture: the Baltimore Hip Studies. J Gerontol A Biol Sci MedSci. 2007;62(12):1402–6.

22. Sokolove J, Lepus CM. Role of inflammation in the pathogenesis ofosteoarthritis: latest findings and interpretations. Ther Adv MusculoskeletDis. 2013;5(2):77–94.

23. Berenbaum F. Osteoarthritis as an inflammatory disease (osteoarthritis is notosteoarthrosis!). Osteoarthritis Cartilage. 2013;21(1):16–21.

24. Nich C, Takakubo Y, Pajarinen J, Ainola M, Salem A, Sillat T, et al.Macrophages—key cells in the response to wear debris from jointreplacements. J Biomed Mater Res A. 2013;101(10):3033–45.

25. Bondeson J, Wainwright SD, Lauder S, Amos N, Hughes CE. The role ofsynovial macrophages and macrophage-produced cytokines in drivingaggrecanases, matrix metalloproteinases, and other destructive andinflammatory responses in osteoarthritis. Arthritis Res Ther. 2006;8(6):R187.

26. Pelham CJ, Agrawal DK. Emerging roles for triggering receptor expressedon myeloid cells receptor family signaling in inflammatory diseases. ExpertRev Clin Immunol. 2014;10(2):243–56.

27. Turnbull IR, Gilfillan S, Cella M, Aoshi T, Miller M, Piccio L, et al. Cutting edge:TREM-2 attenuates macrophage activation. J Immunol. 2006;177(6):3520–4.

28. Takahashi K, Prinz M, Stagi M, Chechneva O, Neumann H. TREM2-transducedmyeloid precursors mediate nervous tissue debris clearance and facilitaterecovery in an animal model of multiple sclerosis. PLoS Med. 2007;4(4):e124.

29. Murakami Y, Akahoshi T, Aoki N, Toyomoto M, Miyasaka N, Kohsaka H.Intervention of an inflammation amplifier, triggering receptor expressed onmyeloid cells 1, for treatment of autoimmune arthritis. Arthritis Rheum.2009;60(6):1615–23.

30. Martel-Pelletier J, Raynauld JP, Dorais M, Abram F, Pelletier JP. The levels ofthe adipokines adipsin and leptin are associated with knee osteoarthritisprogression as assessed by MRI and incidence of total knee replacement insymptomatic osteoarthritis patients: a post hoc analysis. Rheumatology(Oxford). 2015;55(4):680–8.

31. Wang Y, Xu Z, Wang J, Xu S. DUSP19, a downstream effector of leptin,inhibits chondrocyte apoptosis via dephosphorylating JNK duringosteoarthritis pathogenesis. Mol Biosyst. 2016;12(3):721–8.

32. Kandahari AM, Yang X, Dighe AS, Pan D, Cui Q. Recognition of immuneresponse for the early diagnosis and treatment of osteoarthritis. J ImmunolRes. 2015;2015:192415.

33. Gupta GK, Agrawal T, DelCore MG, Mohiuddin SM, Agrawal DK. Vitamin Ddeficiency induces cardiac hypertrophy and inflammation in epicardialadipose tissue in hypercholesterolemic swine. Exp Mol Pathol.2012;93(1):82–90.

34. Ridker PM. Hyperlipidemia as an instigator of inflammation: inauguratingnew approaches to vascular prevention. J Am Heart Assoc. 2012;1(1):3–5.

35. Glushakova O, Kosugi T, Roncal C, Mu W, Heinig M, Cirillo P, et al. Fructoseinduces the inflammatory molecule ICAM-1 in endothelial cells. J Am SocNephrol. 2008;19(9):1712–20.

36. Ronken S, Arnold MP, Ardura Garcia H, Jeger A, Daniels AU, Wirz D. Acomparison of healthy human and swine articular cartilage dynamicindentation mechanics. Biomech Model Mechanobiol. 2012;11(5):631–9.

37. Chu CR, Szczodry M, Bruno S. Animal models for cartilage regeneration andrepair. Tissue Eng Part B Rev. 2010;16(1):105–15.

38. Holick MF. Vitamin D, deficiency. N Engl J Med. 2007;357(3):266–81.39. Hathcock JN, Shao A, Vieth R, Heaney R. Risk assessment for vitamin D.

Am J Clin Nutr. 2007;85(1):6–18.40. Sun Y, Mauerhan DR, Kneisl JS, Norton HJ, Zinchenko N, Ingram J, et al.

Histological examination of collagen and proteoglycan changes inosteoarthritic menisci. Open Rheumatol J. 2012;6:24–32.

41. Houston DK, Tooze JA, Neiberg RH, Hausman DB, Johnson MA, Cauley JA,et al. 25-hydroxyvitamin D status and change in physical performance and

Rai et al. Arthritis Research & Therapy (2016) 18:203 Page 16 of 17

strength in older adults: the Health, Aging, and Body Composition Study.Am J Epidemiol. 2012;176(11):1025–34.

42. Slemenda C, Brandt KD, Heilman DK, Mazzuca S, Braunstein EM, Katz BP,et al. Quadriceps weakness and osteoarthritis of the knee. Ann Intern Med.1997;127(2):97–104.

43. Santangelo KS, Radakovich LB, Fouts J, Foster MT. Pathophysiology ofobesity on knee joint homeostasis: contributions of the infrapatellar fat pad.Horm Mol Biol Clin Investig. 2016;26(2):97–108.

44. Tsavalas N, Karantanas AH. Suprapatellar fat-pad mass effect: MRI findingsand correlation with anterior knee pain. AJR Am J Roentgenol.2013;200(3):W291–6.

45. Zagotta I, Dimova EY, Debatin KM, Wabitsch M, Kietzmann T, Fischer-Posovszky P.Obesity and inflammation: reduced cytokine expression due to resveratrol in ahuman in vitro model of inflamed adipose tissue. Front Pharmacol. 2015;6:79.

46. Meyer M, Sellam J, Fellahi S, Kotti S, Bastard JP, Meyer O, et al. Serum levelof adiponectin is a surrogate independent biomarker of radiographicdisease progression in early rheumatoid arthritis: results from the ESPOIRcohort. Arthritis Res Ther. 2013;15(6):R210.

47. Senolt L, Pavelka K, Housa D, Haluzík M. Increased adiponectin is negativelylinked to the local inflammatory process in patients with rheumatoidarthritis. Cytokine. 2006;35(5–6):247–52.

48. Merlino LA, Curtis J, Mikuls TR, Cerhan JR, Criswell LA, Saag KG. Vitamin Dintake is inversely associated with rheumatoid arthritis: results from the IowaWomen’s Health Study. Arthritis Rheum. 2004;50(1):72–7.

49. Loi F, Cordova LA, Zhang R, Pajarinen J, Lin TH, Goodman SB, et al. Theeffects of immunomodulation by macrophage subsets on osteogenesis invitro. Stem Cell Res Ther. 2016;7(1):15.

50. Butterfield TA, Best TM, Merrick MA. The dual roles of neutrophils andmacrophages in inflammation: a critical balance between tissue damageand repair. J Athl Train. 2006;41(4):457–65.

51. Maskrey BH, Megson IL, Whitfield PD, Rossi AG. Mechanisms of resolution ofinflammation: a focus on cardiovascular disease. Arterioscler Thromb VascBiol. 2011;31(5):1001–6.

52. Di Rosa M, Malaguarnera M, Nicoletti F, Malaguarnera L. Vitamin D3: ahelpful immuno-modulator. Immunology. 2011;134(2):123–39.

53. Goldsmith JR. Vitamin D, as an immunomodulator: risks with deficienciesand benefits of supplementation. Healthcare (Basel). 2015;3(2):219–32.

54. Fernandes de Abreu DA, Eyles D, Féron F. Vitamin D, a neuro-immunomodulator: implications for neurodegenerative and autoimmunediseases. Psychoneuroendocrinology. 2009;34 Suppl 1:S265–77.

55. Agrawal T, Gupta GK, Agrawal DK. Vitamin D deficiency decreases theexpression of VDR and prohibitin in the lungs of mice with allergic airwayinflammation. Exp Mol Pathol. 2012;93(1):74–81.

56. Davies-Tuck ML, Wluka AE, Wang Y, Teichtahl AJ, Jones G, Ding C, et al. Thenatural history of cartilage defects in people with knee osteoarthritis.Osteoarthritis Cartilage. 2008;16(3):337–42.

57. Kleemann RU, Krocker D, Cedraro A, Tuischer J, Duda GN. Altered cartilagemechanics and histology in knee osteoarthritis: relation to clinicalassessment (ICRS Grade). Osteoarthritis Cartilage. 2005;13(11):958–63.

58. Hoshiyama Y, Otsuki S, Oda S, Kurokawa Y, Nakajima M, Jotoku T, et al.Chondrocyte clusters adjacent to sites of cartilage degeneration havecharacteristics of progenitor cells. J Orthop Res. 2015;33(4):548–55.

59. Lotz MK, Otsuki S, Grogan SP, Sah R, Terkeltaub R, D’Lima D. Cartilage cellclusters. Arthritis Rheum. 2010;62(8):2206–18.

60. Dean DD, Schwartz Z, Schmitz J, Muniz OE, Lu Y, Calderon F, et al. Vitamin Dregulation of metalloproteinase activity in matrix vesicles. Connect Tissue Res.1996;35(1–4):331–6.

61. Carrino DA, Lidor C, Edelstein S, Caplan AI. Proteoglycan synthesis in vitaminD-deficient cartilage: recovery from vitamin D deficiency. Connect TissueRes. 1989;19(2–4):135–47.

62. Horton Jr WE, Balakir R, Precht P, Liang CT. 1,25-Dihydroxyvitamin D3

down-regulates aggrecan proteoglycan expression in immortalized ratchondrocytes through a post-transcriptional mechanism. J Biol Chem.1991;266(36):24804–8.

63. Tetlow LC, Woolley DE. The effects of 1α,25-dihydroxyvitamin D3 on matrixmetalloproteinase and prostaglandin E2 production by cells of therheumatoid lesion. Arthritis Res Ther. 1999;1(1):63.

64. Felson DT, Niu J, Clancy M, Aliabadi P, Sack B, Guermazi A, et al. Low levelsof vitamin D and worsening of knee osteoarthritis: results of twolongitudinal studies. Arthritis Rheum. 2007;56(1):129–36.

65. Konstari S, Kaila-Kangas L, Jaaskelainen T, Heliovaara M, Rissanen H,Marniemi J, et al. Serum 25-hydroxyvitamin D and the risk of knee and hiposteoarthritis leading to hospitalization: a cohort study of 5274 Finns.Rheumatology (Oxford). 2014;53(10):1778–82.

• We accept pre-submission inquiries

• Our selector tool helps you to find the most relevant journal

• We provide round the clock customer support

• Convenient online submission

• Thorough peer review

• Inclusion in PubMed and all major indexing services

• Maximum visibility for your research

Submit your manuscript atwww.biomedcentral.com/submit

Submit your next manuscript to BioMed Central and we will help you at every step:

Rai et al. Arthritis Research & Therapy (2016) 18:203 Page 17 of 17